Text Correction System Using Expanded Text for Speech Recognition Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Speech recognition technologies often produce imperfect text transcriptions, necessitating correction methods that can accurately identify and rectify errors while providing reliable alternative text with confidence.
Innovation Solution
A method involving a computing device that determines problem text within text information, generates alternative text using expanded or non-text type information, and provides the alternative text along with its source, utilizing techniques like masking and multiple model evaluations for high accuracy and user convenience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If speech recognition is used to transcribe user speech, then text information can be generated quickly, but the accuracy of the transcribed text deteriorates
Solution Approach 1:
The patent introduces an intermediary correction system that includes a problem text determination unit, alternative text generation unit, and alternative text determination unit. This intermediary layer processes the output from speech recognition, identifies errors, and generates corrected text, thereby maintaining the speed advantage of speech recognition while improving transcription accuracy through systematic error correction
Solution Approach 2:
The system implements feedback mechanisms where the alternative text determination unit evaluates multiple candidate corrections and selects the most appropriate one. The system also provides feedback to users about confidence levels and alternative interpretations, allowing for iterative refinement of transcription accuracy while maintaining efficient processing
2Measurement precision
If multiple models and data search are used to generate alternative text, then correction accuracy is improved, but system complexity increases
Solution Approach 1:
The correction system is segmented into distinct functional units: problem text determination unit, alternative text generation unit, and alternative text determination unit. Each unit handles a specific aspect of the correction process, allowing for modular implementation and management of complexity while achieving high correction accuracy through coordinated operation of these specialized components
Solution Approach 2:
The system dynamically adjusts parameters such as confidence thresholds, model selection, and search depth based on the specific context and problem type. This allows the system to optimize correction accuracy for different scenarios while managing computational resources efficiently, preventing unnecessary complexity from being applied uniformly across all cases
3Reliability
If alternative text with source information is provided, then user confidence is enhanced, but information processing volume increases
Solution Approach 1:
The system applies local quality enhancement by providing source information and confidence metrics specifically for the corrected text segments rather than uniformly for all text. This targeted approach enhances user confidence where it is most needed (in corrected portions) while minimizing unnecessary information volume in already accurate segments
Data Source
AI summary
Disclosed is a method of correcting text information. The method can be performed by a computing device. The method includes obtaining the text information. The method includes determining problem text within the text information. The method includes generating alternative text to correct the problem text by utilizing expanded text associated with the problem text or non-text type information associated with the problem text. The method includes providing information about the alternative text for correcting the problem text.